Do individuals prefer to compete fairly, or unfairly with an opponent? We study individuals who can choose how to compete for one ex-post nonzero payoff. They can either nudge themselves into a fair set of rules where they have the same information and actions as their opponent, or into unfair rules where they spy, sabotage or fabricate their opponent's action. In an experiment, we observe significant altruism under rules which allow for fabrication and sabotage, but not under rules which allow for spying. We provide direct evidence that this altruism emanates from an ethical concern purely about the rules of the game. How individuals deal with this concern - whether they nudge themselves into fabrication-free, spying-free, or sabotage-free rules, or whether they assume the power to fabricate or sabotage to compensate their opponent by giving all payoff away - varies along with individuals' attitudes towards power.